deepface

DeepFace is a lightweight face recognition and facial attribute analysis (age, gender, emotion and race) framework for python. It is a hybrid face recognition framework wrapping state-of-the-art models: VGG-Face, FaceNet, OpenFace, DeepFace, DeepID, ArcFace, Dlib, SFace, GhostFaceNet, Buffalo_L.

A modern face recognition pipeline consists of 5 common stages: detect, align, normalize, represent and verify. While DeepFace handles all these common stages in the background, you don’t need to acquire in-depth knowledge about all the processes behind it. You can just call its verification, find or analysis function with a single line of code.

Experiments show that human beings have 97.53% accuracy on facial recognition tasks whereas those models already reached and passed that accuracy level.

Installation

The easiest way to install deepface is to download it from PyPI. It's going to install the library itself and its prerequisites as well.

$ pip install deepface

Alternatively, you can also install deepface from its source code. Source code may have new features not published in pip release yet.

$ git clone https://github.com/serengil/deepface.git
$ cd deepface
$ pip install -e .

Once you installed the library, then you will be able to import it and use its functionalities.

from deepface import DeepFace

Face Verification - Demo

This function determines whether two facial images belong to the same person or to different individuals. The function returns a dictionary, where the key of interest is verified: True indicates the images are of the same person, while False means they are of different people.

result: dict = DeepFace.verify(img1_path = "img1.jpg", img2_path = "img2.jpg")

Face recognition - Tutorial, Demo

Face recognition requires applying face verification many times. DeepFace provides an out-of-the-box find function that searches for the identity of an input image within a specified database path.

dfs: List[pd.DataFrame] = DeepFace.find(img_path = "img1.jpg", db_path = "C:/my_db")

Here, the find function relies on a directory-based face datastore and stores embeddings on disk. Alternatively, DeepFace provides a database-backed search functionality where embeddings are explicitly registered and queried. Currently, postgres, mongo, neo4j and weaviate are supported as backend databases.

# register an image into the database
DeepFace.register(img = "img1.jpg")

# perform exact search
dfs: List[pd.DataFrame] = DeepFace.search(img = "target.jpg")

If you want to perform approximate nearest neighbor search instead of exact search to achieve faster results on large-scale databases, you can build an index beforehand and explicitly enable ANN search. Here, Faiss is used to index embeddings in postgres and mongo, whereas weaviate and neo4j handle indexing internally.

# build index on registered embeddings (for postgres and mongo only)
DeepFace.build_index()

# perform approximate nearest neighbor search
dfs: List[pd.DataFrame] = DeepFace.search(img = "target.jpg", search_method = "ann")

Facial Attribute Analysis - Demo

DeepFace also comes with a strong facial attribute analysis module including age, gender, facial expression (including angry, fear, neutral, sad, disgust, happy and surprise) and race (including asian, white, middle eastern, indian, latino and black) predictions.

objs: List[dict] = DeepFace.analyze(
  img_path = "img4.jpg", actions = ['age', 'gender', 'race', 'emotion']
)

Age model got ± 4.65 MAE; gender model got 97.44% accuracy, 96.29% precision and 95.05% recall as mentioned in its tutorial.

Real Time Analysis - Demo, React Demo part-i, React Demo part-ii

You can run deepface for real time videos as well. Stream function will access your webcam and apply both face recognition and facial attribute analysis. The function starts to analyze a frame if it can focus a face sequentially 5 frames. Then, it shows results 5 seconds.

DeepFace.stream(db_path = "C:/database")

Even though face recognition is based on one-shot learning, you can use multiple face pictures of a person as well. You should rearrange your directory structure as illustrated below.

user
├── database
│   ├── Alice
│   │   ├── Alice1.jpg
│   │   ├── Alice2.jpg
│   ├── Bob
│   │   ├── Bob.jpg

If you intend to perform face verification or analysis tasks directly from your browser, deepface-react-ui is a separate repository built using ReactJS depending on deepface api.

Here, you can also find some real time demos for various facial recognition models:

Embeddings - Tutorial, Demo

Face recognition models basically represent facial images as multi-dimensional vectors. Sometimes, you need those embedding vectors directly. DeepFace comes with a dedicated representation function.

embedding_objs: List[dict] = DeepFace.represent(img_path = "img.jpg")

Embeddings can be plotted as below. Each slot is corresponding to a dimension value and dimension value is emphasized with colors. Similar to 2D barcodes, vertical dimension stores no information in the illustration.

In summary, the distance between vector embeddings of the same person should be smaller than that between embeddings of different people. When reduced to two-dimensional space, the clusters become clearly distinguishable.

Face recognition models - Demo

DeepFace is a hybrid face recognition package. It currently wraps many state-of-the-art face recognition models: VGG-Face , FaceNet, OpenFace, DeepFace, DeepID, ArcFace, Dlib, SFace, GhostFaceNet and Buffalo_L. The default configuration uses VGG-Face model.

models = [
    "VGG-Face", "Facenet", "Facenet512", "OpenFace", "DeepFace",
    "DeepID", "ArcFace", "Dlib", "SFace", "GhostFaceNet",
    "Buffalo_L",
]

result = DeepFace.verify(
  img1_path = "img1.jpg", img2_path = "img2.jpg", model_name = models[0]
)

dfs = DeepFace.find(
  img_path = "img1.jpg", db_path = "C:/my_db", model_name = models[1]
)

embeddings = DeepFace.represent(
  img_path = "img.jpg", model_name = models[2]
)

FaceNet, VGG-Face, ArcFace and Dlib are overperforming ones based on experiments - see BENCHMARKS for more details. You can find the measured scores of various models in DeepFace and the reported scores from their original studies in the following table.

Conducting experiments with those models within DeepFace may reveal disparities compared to the original studies, owing to the adoption of distinct detection or normalization techniques. Furthermore, some models have been released solely with their backbones, lacking pre-trained weights. Thus, we are utilizing their re-implementations instead of the original pre-trained weights.

Face Detection and Alignment - Demo

Face detection and alignment are important early stages of a modern face recognition pipeline. Experiments show that detection increases the face recognition accuracy up to 42%, while alignment increases it up to 6%. OpenCV, Ssd, Dlib, MtCnn, Faster MtCnn, RetinaFace, MediaPipe, Yolo, YuNet and CenterFace detectors are wrapped in deepface.

All deepface functions accept optional detector backend and align input arguments. You can switch among those detectors and alignment modes with these arguments. OpenCV is the default detector and alignment is on by default.

backends = [
    'opencv', 'ssd', 'dlib', 'mtcnn', 'fastmtcnn',
    'retinaface', 'mediapipe', 'yolov8n', 'yolov8m', 
    'yolov8l', 'yolov11n', 'yolov11s', 'yolov11m',
    'yolov11l', 'yolov12n', 'yolov12s', 'yolov12m',
    'yolov12l', 'yunet', 'centerface',
]
detector = backends[3]
align = True

obj = DeepFace.verify(
  img1_path = "img1.jpg", img2_path = "img2.jpg", detector_backend = detector, align = align
)

dfs = DeepFace.find(
  img_path = "img.jpg", db_path = "my_db", detector_backend = detector, align = align
)

embedding_objs = DeepFace.represent(
  img_path = "img.jpg", detector_backend = detector, align = align
)

demographies = DeepFace.analyze(
  img_path = "img4.jpg", detector_backend = detector, align = align
)

face_objs = DeepFace.extract_faces(
  img_path = "img.jpg", detector_backend = detector, align = align
)

Face recognition models are actually CNN models and they expect standard sized inputs. So, resizing is required before representation. To avoid deformation, deepface adds black padding pixels according to the target size argument after detection and alignment.

RetinaFace and MtCnn seem to overperform in detection and alignment stages but they are much slower. If the speed of your pipeline is more important, then you should use opencv or ssd. On the other hand, if you consider the accuracy, then you should use retinaface or mtcnn.

The performance of RetinaFace is very satisfactory even in the crowd as seen in the following illustration. Besides, it comes with an incredible facial landmark detection performance. Highlighted red points show some facial landmarks such as eyes, nose and mouth. That's why, alignment score of RetinaFace is high as well.

The Yellow Angels - Fenerbahce Women's Volleyball Team

You can find out more about RetinaFace on this repo.

Face Anti Spoofing - Demo

DeepFace also includes an anti-spoofing analysis module to understand given image is real or fake. To activate this feature, set the anti_spoofing argument to True in any DeepFace tasks.

# anti spoofing test in face detection
face_objs = DeepFace.extract_faces(img_path="dataset/img1.jpg", anti_spoofing = True)
assert all(face_obj["is_real"] is True for face_obj in face_objs)

# anti spoofing test in real time analysis
DeepFace.stream(db_path = "C:/database", anti_spoofing = True)

Similarity - Demo

Face recognition models are regular convolutional neural networks and they are responsible to represent faces as vectors. We expect that a face pair of same person should be more similar than a face pair of different persons.

Similarity could be calculated by different metrics such as Cosine Similarity, Angular Distance, Euclidean Distance or L2 normalized Euclidean. The default configuration uses cosine similarity. According to experiments, no distance metric is overperforming than other.

metrics = ["cosine", "euclidean", "euclidean_l2", "angular"]

result = DeepFace.verify(
  img1_path = "img1.jpg", img2_path = "img2.jpg", distance_metric = metrics[1]
)

dfs = DeepFace.find(
  img_path = "img1.jpg", db_path = "C:/my_db", distance_metric = metrics[2]
)

API - Demo, Docker Demo

DeepFace serves an API as well - see api folder for more details. You can clone deepface source code and run the api with the following command. It will use gunicorn server to get a rest service up. In this way, you can call deepface from an external system such as mobile app or web.

cd scripts && ./service.sh

Alternatively, you can run the dockerized service.

cd scripts && ./dockerize.sh

Face verification, facial attribute analysis, vector representation and register & search functions are covered in the API. The API accepts images as file uploads (via form data), or as exact image paths, URLs, or base64-encoded strings (via either JSON or form data).

$ curl -X POST http://localhost:5005/represent -d '{"model_name":"Facenet", "img":"img1.jpg"}' -H "Content-Type: application/json"
$ curl -X POST http://localhost:5005/verify -d '{"img1":"img1.jpg", "img2":"img3.jpg"}' -H "Content-Type: application/json"
$ curl -X POST http://localhost:5005/analyze -d '{"img": "img2.jpg", "actions": ["age", "gender"]}' -H "Content-Type: application/json"
$ curl -X POST http://localhost:5005/register -d '{"model_name":"Facenet", "img":"img18.jpg"}' -H "Content-Type: application/json"
$ curl -X POST http://localhost:5005/search -d '{"img":"img1.jpg", "model_name":"Facenet"}' -H "Content-Type: application/json"

Here, you can find a postman project to find out how these methods should be called.

Encrypt Embeddings - Demo with PHE, Tutorial for PHE, Demo with FHE, Tutorial for FHE

Vector embeddings, though not reversible, carry sensitive information like fingerprints, making their security crucial. Encrypting them prevents adversarial misuse. Traditional encryption (e.g., AES) is secure but unsuitable for cloud-based distance calculations.

Homomorphic encryption allows computations on encrypted data without revealing content—ideal for secure cloud processing. For example, the cloud can compute encrypted similarity without knowing the data, while only the key holder can decrypt the result. See the LightPHE library for partially homomorphic encryption.

from lightphe import LightPHE

# build an additively homomorphic cryptosystem (e.g. Paillier) on-prem
cs = LightPHE(algorithm_name = "Paillier", precision = 19)

# define encrypted and plain vectors
encrypted_alpha = DeepFace.represent("source.jpg", cryptosystem=cs)[0]["encrypted_embedding"]
beta = DeepFace.represent("target.jpg")[0]["embedding"]

# dot product of encrypted & plain embedding in cloud - private key not required
encrypted_cosine_similarity = encrypted_alpha @ beta

# decrypt similarity on-prem - private key required
calculated_similarity = cs.decrypt(encrypted_cosine_similarity)[0]

# verification
print("same person" if calculated_similarity >= 1 - threshold else "different persons")

For stronger privacy, fully homomorphic encryption enables dot product computations between encrypted embeddings, but it's far more computationally intensive. Explore CipherFace for FHE-based approaches.

Extended Applications

DeepFace can also be used for fun and insightful applications such as

Find Your Celebrity Look-Alike - Demo, Real-Time Demo, Tutorial

DeepFace can analyze your facial features and match them with celebrities, letting you discover which famous personality you resemble the most.

Find Which Parent a Child Look More - Demo, Tutorial

DeepFace can also be used to compare a child's face to their parents' or relatives' faces to determine which one the child resembles more.

Contribution

Pull requests are more than welcome! If you are planning to contribute a large patch, please create an issue first to get any upfront questions or design decisions out of the way first.

Before creating a PR, you should run the unit tests and linting locally by running make test && make lint command. Once a PR sent, GitHub test workflow will be run automatically and unit test and linting jobs will be available in GitHub actions before approval.

Support

There are many ways to support a project - starring⭐️ the GitHub repo is just one 🙏 It really helps the project get discovered by more people.

If you do like this work, then you can support it financially on Patreon, GitHub Sponsors or Buy Me a Coffee.

Citation

Please cite deepface in your publications if it helps your research.

@article{serengil2024lightface,
  title     = {A Benchmark of Facial Recognition Pipelines and Co-Usability Performances of Modules},
  author    = {Serengil, Sefik and Ozpinar, Alper},
  journal   = {Journal of Information Technologies},
  volume    = {17},
  number    = {2},
  pages     = {95-107},
  year      = {2024},
  doi       = {10.17671/gazibtd.1399077},
  url       = {https://dergipark.org.tr/en/pub/gazibtd/issue/84331/1399077},
  publisher = {Gazi University}
}

@inproceedings{serengil2021lightface,
  title        = {HyperExtended LightFace: A Facial Attribute Analysis Framework},
  author       = {Serengil, Sefik Ilkin and Ozpinar, Alper},
  booktitle    = {2021 International Conference on Engineering and Emerging Technologies (ICEET)},
  pages        = {1-4},
  year         = {2021},
  doi          = {10.1109/ICEET53442.2021.9659697},
  url          = {https://ieeexplore.ieee.org/document/9659697},
  organization = {IEEE}
}

@inproceedings{serengil2020lightface,
  title        = {LightFace: A Hybrid Deep Face Recognition Framework},
  author       = {Serengil, Sefik Ilkin and Ozpinar, Alper},
  booktitle    = {2020 Innovations in Intelligent Systems and Applications Conference (ASYU)},
  pages        = {23-27},
  year         = {2020},
  doi          = {10.1109/ASYU50717.2020.9259802},
  url          = {https://ieeexplore.ieee.org/document/9259802},
  organization = {IEEE}
}

Also, if you use deepface in your GitHub projects, please add deepface in the requirements.txt.

Licence

DeepFace is licensed under the MIT License - see LICENSE for more details.

DeepFace wraps some external face recognition models: VGG-Face, Facenet (both 128d and 512d), OpenFace, DeepFace, DeepID, ArcFace, Dlib, SFace, GhostFaceNet and Buffalo_L. Besides, age, gender and race / ethnicity models were trained on the backbone of VGG-Face with transfer learning. Similarly, DeepFace wraps many face detectors: OpenCv, Ssd, Dlib, MtCnn, Fast MtCnn, RetinaFace, MediaPipe, YuNet, Yolo and CenterFace. Finally, DeepFace is optionally using face anti spoofing to determine the given images are real or fake. License types will be inherited when you intend to utilize those models. Please check the license types of those models for production purposes.

DeepFace logo is created by Adrien Coquet and it is licensed under Creative Commons: By Attribution 3.0 License.